Metabolism of Clomipramine in a Japanese Psychiatric Population: Hydroxylation, Desmethylation, and Glucuronidation Kazutaka Shimada, M.D., Ph.D., Toshifumi Noguchi, M.D., Ph.D., Yuhi Ozeki, M.D., Sachiyo Morita, M.D., Morikazu Shibasaki, M.D., Toshiyuki Someya, M.D., and Saburo Takahashi, M.D., Ph.D.

------We measured the concentrations of clomipramine and its positive correlations were observed between these drug metabolites, N-desmethylclomipramine, 8-hydroxy-N­ concentrations and daily doses of clomipramine desmethylclomipramine, 8-hydroxyclomipramine by high­ hydrochloride (mg/kg body weight). Although the performance liquid chromatography in 108 Japanese metabolic ratios for desmethylation, hydroxylation, and psychiatric patients receiving clomipramine hydrochloride glucuronidation that were calculated from steady-state PO. The concentrations of the glucuronide conjugates of drug concentrations varied substantially with 36-, 14-, 8-hydroxyclomipramine and 8-hydroxy-N-desmethyl­ and 28-fold interindividual variations, respectively, clomipramine were assayed via enzymatic hydrolysis. apparent poor desmethylators, poor hydroxylators, or Although there were large interindividual variations of poor glucuronidators were not found. concentrations of parent, intermediate metabolic [Neuropsychopharmacology 12:323-333, 1995] compounds, and glucuronide conjugates, significant ·------·------

KEY WORDS: Tricyclic antidepressants: Clomipramine; TCAs varied considerably between individuals treated Hydroxylation; Desmethylation; Glucuronidation even with similar doses of the same drug, which sug­ Interindividual differences in drug metabolism, which gests that such pharmacokinetic variation reflected in­ determine plasma levels or adverse drug effects, have terindividual differences in drug metabolism (Hammer been observed. In particular, the existence of poor and Sjoqvist 1967). Alexanderson et al. (1969) also in­ metabolizers has been noted for various drugs, includ­ vestigated steady-state plasma levels of nortriptyline ing debrisoquine, sparteine, mephenytoin, or desipra­ in monozygotic and dizygotic twins, concluding that mine (Jacqz et al. 1986). For example, plasma levels of variability in steady-state plasma concentrations be­ tween individuals was mostly genetically determined.

------Clomipramine (C), a chlorinated analog of imipra­ From the Department of Psychiatry (KM, TM, YO, SM, MS, ST), mine, has become one of the most frequently used TCAs Shiga University of Medical Science, Otsu, Shiga, Japan; Psychiatry since 1960s for the treatment of psychiatric illness. Service (TN), Hachiman Sejukai Hospital, Omihachiman, Shiga, Ja­ pan; Psychiatry Service (SM), Toyosato Hospital, Toyosatocho, In­ Clomipramine has been reported to be a potent inhibi­ ukami County, Shiga, Japan; Psychiatry Service (MS), Takashima tor of serotonin reuptake, and its active desmethylated General Hospital, Takashimacho, Takashima County, Japan; Health and/or hydroxylated metabolites inhibit the reuptake Administration Center (TS), Shiga University of Medical Science, Seta Tsukinowacho, Otsu, Shiga, Japan. of both serotonin and/or noradrenaline. Address correspondence to: Kazutaka Shimoda, M.D., Ph.D., Despite the widespread use of C, relatively little is Department of Psychiatry, Shiga University of Medical Science, Seta known about its metabolism and phamacokinetics in Tsukinowacho, Otsu, Shiga, 520-21, Japan. Received May 13, 1994; revised November 8, 1994; accepted Novem­ humans. By analogy with the metabolism of imipra­ ber 11, 1994. mine, the major route of biotransformation of C is des-

NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4 © 1995 American College of Neuropsychopharmacology Published by Elsevier Science Inc. 0893-133X/95/$9 .50 655 Avenue of the Americas, New York, NY 10010 SSDI 0893-133X(94)00098-K 324 K. Shimoda et al. NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4

C DC

HC HOC

Cl N CH2CH2CH2N(CH3h ,r-1 hr\~- Urine~~ Figure 1. Structural formulae of clomipramine {C), N-desmethylclomipramine (DC), 8-hydroxyclomipramine (HC), and 8-hydroxy-N-desmethylclomipramine (HDC). Metabolic pathways of Care also shown. Solid, open, and shaded arrows show hydroxylation clearance, desmethylation clearance, and the combined effect of conjugation and direct renal excretion of hydroxylated metabolites.

methylation at nitrogen of the N-demethylamino group hydroxylation of C, and glucuronidation of HC and (see Figure 1). Other metabolic pathways comprise HOC. hydroxylation clearance and conjugation of hydrox­ To our knowledge, this is the first comprehensive ylated metabolites, as shown in Figure 1. report dealing with desmethylation, hydroxylation, Desmethylclomipramine (DC), 8-hydroxyclomi­ and glucuronidation of TCAs in a pramine (HC), 8-hydroxy-N-desmethylclomipramine sizable Asian population. The present report may be informative when compared (HOC), 8-hydroxyclorr.ipramine glucuronide (HCG), with previous data on metabolism of TCAs and 8-hydroxy-N-desmethylclomipramine glucuronide in other eth­ nic groups. (HDCG), have been found in the plasma of patients who were administered clomipramine hydrochloride (Nielsen et al. 1992). This study was conceived to investigate the follow­ SUBJECTS AND METHODS ing matters: Subjects 1. The plasma concentrations of C, HC, DC, HOC, One hundred and eight patients HCG, and HDCG in psychiatric patients regularly (male: n = 32; female: n = 76) receiving clomipramine receiving clomipramine hydrochloride PO. hydrochloride (ANA­ FRANIL ®, Ciba-Geigy Japan Ltd., Takarazuka, Japan) 2. The impact of age, nutritional state, daily dose of participated in this study. Of these patients 87 patients clomipramine hydrochloride, administration of were hospitalized at the Shiga University of Medical benzodiazepine derivatives, smoking, or gender on Science Hospital, Toyosato Hospital, or Hirosaki Uni­ desmethylation, hydroxylation of C, and glucuroni­ versity Hospital for the treatment of their psychiatric dation of hydroxylated metabolites of C (i.e., HC illnesses, while 21 subjects were under treatment at the and HOC). outpatient clinic of these hospitals. Informed consent 3. Interindividual variations of desmethylation, was obtained from each subject. NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4 Clomipramine Metabolism in an Asian Population 325

Patients on neuroleptics, barbiturates, and/or oral patients were instructed not to take alcohol and also contraceptives as subordinate prescriptions were ex­ instructed to report if they took medications such as cluded because such drugs are known to alter the me­ laxatives, antihistamines, analgesics, or contraceptives. tabolism (Daniel and Melzacka 1986; Kruger et al. 1986) Their family members were entrusted with the care of and/or plasma levels (Alexanderson et al. 1969; Brct,sen the prescribed drugs and instructed to confirm regular et al. 1986a; Edelbroek et al. 1987) of TCAs. Benzodi­ intake of the drugs. Patients were maintained for at least azepines in usual doses were allowed to be ad­ 14 days at the same dosage until the time of blood sam­ ministered as sleep inducers when patients complained pling. Blood (20 ml) was collected at 6 A.M., 12 hours of sleep disturbances since benzodiazepines have been after the evening dose, into heparinized Venoject® tubes reported not to influence substantially the metabolism (Terumo Japan, Tokyo, Japan) and centrifuged at of TCAs (Bertilsson et al. 1988; Otani et al. 1987; Trask­ 1,000 g. Aliquots of plasma were frozen and stored at man et al. 1979). The drugs and their doses were: - 80°C until assayed. Seventy-four additional samples flunitrazepam (ROHYPNOL®, Nippon Roche K.K., were obtained from 52 patients, 2, 4, or 6 weeks after Tokyo, Japan) 1-4 mg/day; nitrazepam (BENZALINE®, the frrst blood sampling, and a total of 182 samples were Shionogi & Co., Osaka, Japan), 5 mg/day; flurazepam collected. (BENOZIL®, Kyowa Hakko Co., Ltd., Tokyo, Japan), 30 mg/day, diazepam (HORIZON®; YamanouchiPhar­ maceutical Co., Ltd., Tokyo, Japan), 5 mg/day; broma­ Determination of Plasma Levels of C and Its zepam (LEXOTAN®, Nippon Roche, Tokyo, Japan), 5 Metabolites DC, Unconjugated HOC mg/day; triazolam (HALCION®, Japan Upjohn Ltd., and Unconjugated HC Tokyo, Japan), 0.25 mg/day; etizolam (DEPAS®, Yo­ Plasma levels of C and its metabolites DC, unconjugated shitomi Pharmaceutical Industries, Ltd., Osaka, Ja­ HOC, unconjugated HC were determined by high­ pan), 1 mg/day; brotizolam (LENDORMIN®, Nippon performance liquid chromatography (HPLC) according Boehinger-Ingelheim, Kawanishi, Japan), 0.25 mg/day, to the procedure of Fekete et al. (1981) with minor haloxazolam (SOMELIN®, Sankyo Co., Ltd., Tokyo, modi&cations. 1 The recoveries of C, DC, HC, and in­ Japan), 10 mg/day; rilmazafone (RHYTHMY®, Shionogi ternal standard were 85% to 90% and approximately & Co., Ltd., Osaka, Japan), 1 mg/day. 50% for HOC. The intraassay coefficients of variation Demographic data, medical histories, and labora­ (CVs) were 2% to 4% (n = 10) for each compound. The tory data including hematology, serology, electrolytes, interassay CVs were 10% to 13% (n = 7) for each com­ and urine analysis were collected for each patient. Pa­ pound. tients with physical illnesses were excluded from this study. The one hundred and eight patients consisted of Deconjugation and Determination of 32 men and 76 women, and their ages ranged from 17 Plasma Levels of HCG and HDCG to 73 (mean± SD= 44.5 ± 14.5) years. Diagnoses were made by semistructured interviews using DSM-III-R HCG and HDCG in plasma were deconjugated by the criteria (American Psychiatric Association 1987). method of Bock et al. (1982) and Bock et al. (1983) with More than 80% of the patients (89/108) were diag­ minor modi&cations. 1 Five hundred microliter of sam­ nosed as having a major depressive episode. Their body ple plasma was incubated with 2,500 U of P-glucuron­ weights and heights were in a range of 32.3 to 80.8 idase (P-D-Glucuronide glucuronosohydrolase from (mean± SD= 52.3 ± 9.5)kgand 136to179.5(157.1 ± Helix pomatia, Type H-2; SIGMA Chemicals, MO, 10.3) cm, respectively, body mass index !BMI = [body USA). The optimum hydrolytic condition, which pro­ weight (kg)]/[height (m)]2] was calculated for each sub­ vided maximum yield of unconjugated metabolites, was ject as an index of his or her nutritional state. determined by the preliminary experiments using pooled plasma that contained 10% of urine from a subject who was administered 25 mg/day IV of clomipramine Blood Sampling and Administration of hy­ drochloride for 14 days. Clomipramine Hydrochloride Because the pure HCG or HDCG was not available, Daily doses of clomipramine hydrochloride were we could not determine the absolute efficiency of the titrated in a range of 30 to 250 (111.8 ±52.3) mg/day, enzymatic hydrolysis. However, interassay CVs (n = or 0.40 to 5.10 (2.18 ± 1.08) mg/kg body weight. 6) for total HC (i.e., HC + HCG) and total HOC (i.e., ANAFRANIL® (Ciba-Geigy Japan Ltd., Takarazuka, Ja­ HOC + HDCG) were 9% (3,707 ± 345 nmol/1) and 3% pan) tablets containing 10 (57 mol) or 25 mg (142 mol) of clomipramine hydrochloride were administered orally with tap water. For inpatients, nurses were in­ 1 Detailed description of analytical procedures should be requested structed to confirm intake of the prescribed drugs. Out- from K. Shimoda. 326 K. Shimoda et al. NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4

(2,954 ± 92 nmol), respectively. Intraassay CVs (n = glucuronidation pathways (i.e., HC->- HCC, HOC->­ 10) for total HC and total HOC were 7% and 7%, respec­ HDCC) are regulated by an identical isozyme. tively.

RESULTS Calculation of Metabolic Ratios for Desmethylation, Hydroxylation, and Glucuronidation Ratio Plasma Levels of C and Its Unconjugated Metabolites Because the main oxidative pathway of C is desmethy­ There were pronounced interindividual variations in lation to DC and plasma concentrations of C and HC the total unconjugated drug concentrations (i.e., C + are strongly dependent on N-desmethylation (Balant­ DC + HC + HOC) and those of each plasma concen­ Gorgia et al. 1986), the C-to-DC ratio was defmed as tration of unconjugated compound; that is, the total un­ the metabolic ratio for desmethylation (MRD), that is, conjugated drug concentrations ranged from 174 to MRD = C/DC, and the DC-to-total HOC ratio [i.e., 8,753 nmol/1 (means ± so = 1,851 ± 1,290 nmol/1), the DC/(HDC + HDCC)] was defined as the metabolic ra­ concentrations of C ranged from 47 to 2,992 nmol/1 tio for hydroxylation (MRH). The ratio of HC + HOC (566 ± 450 nmol/1, those of DC from 29 to 4,242 nmol/1 to HCC + HDCC [i.e., (HC + HDC)/(HCC + HDCC)] (698 ± 604 nmol/1), of HC from 39 to 983 nmol (249 .3 ± was defined as the glucuronidation ratio (GR) because 147.0nmol/l), andofHDCfromOto 1,012nmol/l (339 ± a significantly positive correlation was observed be­ 225 nmol/1). Significant positive correlations were ob­ tween HC/HCG and HDC/HDCG (n = 182; df = 180; served between total unconjugated drug concentrations r = 0.48; p < .001), which means that these two (r = 0. 79; n = 182; df = 180; p < .001) or concentrations

3000~------~----, ~4000 ::: 2000 0 0--- E D O E C 0 D C 2000 G 1000 u 0

2 4 6 2 4 6 Daily dose/body weight (mg/kg) Daily dose/body weight (mg/kg)

1000 1000

D --- ooD DD 0--- 0 D D D E [J 0 o on41 E D 500 0 C r. clg ,crfr b .s .s 500 Li iJiCJO~ Q ~n D cc g_,cJio D D D ':Jk DU 0 01 u u 0 J' [§)- ri~~ n: sflf'i)(h 0 n 'b o I Acff'i~o o o rJ o< ' I l?J J O o ;r; ~I ~ll D c®~1 0 ::::iS n 0 n' n a 0 0 0 2 4 6 0 2 4 6 Oaily dose/body weight (mg/kg) Daily dose/body weight (mg/kg)

2000~------, :; 2000 C D DC' 0 ---0 E E C ..S 1000 CJ 8 1000 u 0 I I

0 2 4 6 2 4 6 Daily dose/body weight (mg/kg) Daily dose/body weight (mg/kg)

Figure 2. A significant and positive correlation between the daily dose of clomipramine hydrochloride per body weight and plasma levels of the parent compound, its desmethylated metabolites, unconjugated hydroxy lated metabolites, or con­ jugated hydroxylated metabolites (C: clomipramine, DC: N-desmethylclomipramine, HC: 8-hydroxyclomipramine, HDC: 8-hydroxy-N-desmethylclomipramine, HCG: 8-hydroxyclomipramine glucuronide, HDCG: 8-hydroxy-N-desmethylclo­ mipramine glucuronide). Coefficients of correlation are 0.61, 0.78, 0.70, 0.74, 0.65, and 0.65 for C, DC, HC, HDC, HCG, and HDCG, respectively (n = 182; df = 180; p < .001, respectively). NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4 Clomipramine Metabolism in an Asian Population 327

of each unconjugated compound in plasma and daily methylation, hydroxylation, and glucuronidation in doses of clomipramine hydrochloride (mg/kg body each subject were determined, and the impact of age, weight) (r = 0.61, 0.78, 0.70, and 0.74 for C, DC, HC, nutritional state ( = BMI), daily dose of clomipramine and HDC; n = 182; df = 180; p < .001, respectively; hydrochloride, administration of benzodiazepine de­ see Figure 2). rivatives, smoking, gender, or age on desmethylation, hydroxylation of clomipramine, and glucuronidation Plasma Levels of Conjugated HCG and HDCG of hydroxy lated metabolites of clomipramine were ana­ Incubation of each sample with ~-glucuronidase before lyzed. extraction did not increase plasma concentrations of DC MRD. Significant relationship between daily dose of and C (average factor = 0.98 and 0.97, respectively), clomipramine (mg/kg body weight) and MRD was de­ while the deconjugation of each sample increased the tected (r = 0.31; df = 180; p < .001). The mean CV of plasma concentrations of HC and HDC by average fac­ MRD, representative of intraindividual variation, was tors of 3.36 and 3.61, respectively. 18.4%. The MRDs were averaged in each of 52 patients Plasma concentrations of HCG and HDCG were whose blood samples were taken more than twice in determined by total HC and HDC (concentrations af­ order to calculate the individual MRD. No significant ter hydrolysis) minus unconjugated HC or HDC (con­ relationship was observed between age of subjects and centrations without hydrolysis), respectively. There individual MRD (r = 0.08; n = 108; df = 106; NS). No sig­ were pronounced interindividual variations in the con­ nificant relationship between BMI and individual MRD jugated hydroxylated drug concentrations (i.e., HCG was detected either (r = 0.07; n = 108; df = 106; NS). and HDCG). Concentrations of HCG were 45 to 1,855 A three-way analysis of variance (three-way nmol/1 (561 ± 381 nmol/1); of HDCG, 64 to 2,464 nmol/1 ANOVA) revealed that gender and smoking had (676 ± 450 nrnol/1); while significant positive correlations significant effects on MRDs while administration of ben­ were observed between conjugated hydroxylated drug zodiazepines had no significant effect at all (Table 1). concentrations (HCG, r = 0.65; HDCG, r = 0.65, n = 182, df = 180, p < .001, respectively) and the daily dose MRH. Significant relationship between daily dose of of clomipramine (mg/kg body weight) as shown in Fig­ clomipramine (mg/kg body weight) and MRH was ob­ ure 2. served (r = 0.53; df = 180; p < .001). The mean CV of MRH, representative of intraindividual variation, was 15.6%. The MRDs were averaged in each of 52 patients Impact of Age, Nutritional State, Daily Dose of whose blood samples were taken more than twice in or­ Clomipramine Hydrochloride, Administration of der to calculate the individual MRH. No significant rela­ Benzodiazepine Derivatives, Smoking or Gender on tionship was observed between age of subjects and in­ Desmethylation, Hydroxylation of C and dividual MRH (r = 0.14; n = 108; df = 106; NS). No Glucuronidation of Hydroxylated Metabolites of C significant relationship between BMI and individual MRH From plasma concentrations of C, DC, HC, HDC, HCG, was detected either (r = 0.08; n = 108; df = 106; NS). and HDCG the metabolic ratios representative of des- A three-way ANOVA revealed that gender and smok-

Table 1. Results of Analysis of Effect of Gender, Benzodiazepines, and Smoking on MRD by Three-way ANOVA Source of Sum of Mean Variation Square df Square F p Main effects 7.96 3 2.65 2.69 0.050 Gender 6.07 1 6.07 6.16 O.DlS Benzodiazepine 0.56 1 0.56 0.57 0.454 Smoking 3.98 1 3.98 4.04 0.047

Table 2. Results of Analysis of Effect of Gender, Benzodiazepines, and Smoking on MRH by Three-way ANOVA Source of Sum of Mean Variation Square df Square F p Main effects 0.70 3 0.23 2.21 0.091 Gender 0.46 1 0.46 4.40 0.039 Benzodiazepine 0.02 1 0.02 0.14 0.707 Smoking 0.51 1 0.51 4.87 0.030 328 K. Shimoda et al. NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4

Table 3. Results of Analysis of Effect of Gender, Benzodiazepines, and Smoking on GR by Three-way ANOV A Source of Sum of Mean Variation Square df Square F p Main effects 3.16 3 1.05 5.79 0.010 Gender 0.69 1 0.69 3.78 0.055 Benzodiazepine 1.51 1 1.51 8.30 0.005 Smoking 0.18 1 0.18 0.98 0.325

ing had signilicant effects on MRHs while administration patterns were linear, which meant that the MRH data ofbenzodiazepines had no signilicant effect at all (Table 2). had a normal distribution pattern. GR. No significant relationship between daily dose of clomipramine (mg/kg body weight) and GR was observed Interindividual Variation of GR (r = 0.01; df = 180; NS). The mean CV of GR, represen­ The individual GR exhibit pronounced interindividual tative of intraindividual variation, was 16%. The GRs were difference, ranging from 0.13 to 3.69 (0.56 ± 0.45). Fig­ averaged in each of 52 patients whose blood samples were ure 5 shows the distribution of individual GR. Approx­ taken more than twice in order to calculate the individual imately 28-fold interindividual variations (0.13-3.69) GR. No significant relationship was observed between were observed in GR. Each value was subjected to nat­ age of subjects and individual GR (r = 0.03; n = 108; df = ural logarithmic transformations in order to draw fre­ 106; NS). No signilicant relationship between BMI and quency distribution histograms of GR as shown in the individual GR was detected either (r = 0.15; n = 108; upper part of Figure 5. Ln(GR) ranging from ca. - 2.0 df = 106; NS). to 1.3 appeared to have a normal distribution pattern. A three-way ANOVA revealed that administration A probit plot of Ln(GR) (lower part of Figure 5) showed of benzodiazepines had signilicant effects on GRs while effect of gender on GRs reached almost signilicance level (p = 0.055, Table 3). 6 lnterindividual Variations of Individual ...... (j) MRD and MRH (.) -~4 ..0 The individual metabolic ratios representative of des­ :J methylation and hydroxylation (MRD and MRH, (j) '+- respectively) varied substantially, ranging from 0.25 to ~2 9.06 (1.17 ± 1.02) and from 0.14 to 2.04 (0.63 ± 0.32), 0 respectively. z Figure 3 shows the distribution of individual meta­ bolic ratio for desmethylation of C (i.e., MRD). Approx­ 0 imately 36-fold interindividual variations (0.25-9.06) 8 were observed in the MRD. Each value was subjected D to natural logarithmic transformations in order to draw D frequency distribution histograms of the MRD, as shown in the upper part of Figure 3. Ln(MRD) ranging ...... 6 15 from about -1.5 to 2.25 appeared to have a normal dis­ 0 tribution pattern. A probit plot of Ln(MRD) (lower part I,_ Q.. 4 of Figure 3) gave a straight line, which is consistent with a normal distribution. f Figure 4 shows the distribution of individual meta­ D bolic ratio for hydroxylation of C (i.e., MRH). Approxi­ 2 mately 14-fold interindividual variations of MRH, rang­ -2 0 2 ing from 0.14 to 2.04 were observed. Ln(MRH) ranging Ln(C/DC) from -2.0 to 0.75 appeared to have a normal distribu­ Figure 3. Frequency histogram (top) and probit plot, or cu­ tion pattern. When the MRH data were transformed mulative frequency distribution (bottom) of MRD (i.e., C/DC) into a probit plot (lower part of Figure 4), the probit plot in 108 subjects studied. NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4 Clomipramine Metabolism in an Asian Population 329

8 8

(/) (/) +-' ..... (.) 6 (.) 6- Q) Q) 75 15 ~4 ~4 ...... 0 0 02 ci 2 z z 0 0

8 8 D D D D D D (§1 D ifP§i ..... 6 ..... 6 :.a :.a 0 0 I,... L. a.. 4 ,/ a..4 0/ D D 2 2 -2 -1 0 1 -2 0 2 Ln[DC/(HDC+HDCG)] Ln[(HC+HDC)/(HCG+HDCG)I Figure 4. Frequency histogram (top) and probit plot, or cu­ Figure 5. Frequency histogram (top) and probit plot, or cu­ mulative frequency distribution (bottom) of MRH [i.e., DC/ mulative frequency distribution (bottom) of GR (i.e., [(HC + (HDC + HDCG)] in 108 subjects studied. HDC)/(HCG + HDCG)] in 108 subjects studied.

a straight line, which meant GRs had a normal distri­ probe drugs, such as debrisoquine, sparteine, or me­ bution. phenytoin, is informative in this kind of study, because polymorphism of oxidation of some drugs, including sparteine and debrisoquine, has been reported to be DISCUSSION associated with interindividual variation in hydroxyla­ tion of various TCAs: amitriptyline (Balant-Gorgia et The present study was designed to know plasma con­ al. 1982; Mellstrom et al. 1983, 1986), nortriptyline (Mell­ centrations of C, DC, HC, HOC, HCC, and HDCG and strom et al. 1981; Nordin et al. 1985; Woolhouse et al. to explore the impact of various factors that have been 1984); imipramine (Brq,sen et al. 1986b) and desipramine reported to influence the metabolism of TCAs such as (Bertilsson and Aberg-Wistedt 1983; Brq,sen et al. 1986b; age, nutritional state, daily dose of the drug, adminis­ Spina et al. 1984, 1987). In addition, desmethylation of tration of benzodiazepine derivatives, smoking, or gen­ imipramine has been reported to be partly mediated der on C metabolism. We also tried to investigate the by mephenytoin oxygenase (Gram et al. 1990; Skjelbo extent of interindividual variations of C metabolism by et al. 1991), which means that poor metabolizers of determining individual metabolic ratios for hydroxyla­ mephenytoin cannot desmethylate imipramine effi­ tion, desmethylation, and glucuronidation from the ciently. However, such probe drugs have not been ap­ steady-state plasma levels of C and its metabolites, in­ proved by the Japanese Ministry of Health and Wel­ cluding glucuronide conjugates. fare for clinical use, and the administration of such This design provides various advantages such as drugs to patients poses an ethical problem in Japan. the easy availability of the data. It can benefit patients Dextromethorphan is available as an over-the-counter in ordinary clinical settings, involves minimal interfer­ drug in our country and is used as one of the substi­ ence with treatment planning for patients, and enables tutes for probe drugs such as debrisoquine, sparteine, the collection of sizable data and therapeutic drug or mephenytoin. Genotyping using the molecular bio­ monitoring simultaneously. logical technique is also an alternative for such probe Of course, simultaneous phenotyping by using drug tests. 330 K. Shimoda et al. NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4

Plasma Levels of C, DC, HC, HOC HDCG/HDC = 2.61), which is 50% less than the results from Bock et al. (1982), Nielsen et al. (1992), and Suftin Della Corte et al. (1979) reported a significant correla­ et al. (1988). tion between the daily dose of clomipramine hydrochlo­ Although one of the possible explanations for this ride per body weight and plasma levels of C and DC. discrepancy could be the low efficiency of hydrolysis Consistent with that study, we obtained data that sug­ used in this study, the condition used in this study was gested strong positive correlations between plasma con­ confirmed to give sufficiently high efficiency of enzy­ centrations of each unconjugated compound and daily matic hydrolysis by preliminary experiments. In addi­ doses of clomipramine hydrochloride per body weight tion, hydrolysis of reference samples (pooled plasma (see Figure 2), thus indicating that the daily dose of that contained 10% of urine from a subject who was clomipramine hydrochloride per body weight is an im­ administered 25 mg per day of clomipramine hydro­ portant determinant of plasma levels of C and its me­ tabolites. chloride intravenously for 14 days) increased plasma concentrations of HC (up to ca. 3,700 nmol/1) and of HDC (up to ca. 3,000 nmol/1) by average factors of ca. Plasma Concentrations of HCG and HDCG 14 and 10, respectively, which indicates the high efficiency of our method for enzymatic hydrolysis. Incubation of each sample with B-glucuronidase before A second possible reason for this discrepancy is extraction did not increase plasma concentration of DC substrate specificity of isozyme of glucuronosyltrans­ and C, which means that glucuronide conjugate of C ferase. Glucuronidation is catalyzed by a family of or DC could not be detected by at least the assay sys­ closely related isozymes that have different but over­ tem used in this study. On the contrary, plasma con­ lapping substrate specificities. The existence of multi­ centrations of HC and HDC were increased after enzy­ ple forms of this enzyme has been suggested (Siest et matic hydrolysis, which means that hydroxylated al. 1987). metabolites of C are exclusively subjected to conjuga­ A third possible reason is the existence of an in­ tion. These findings are consistent with the report of terethnic difference in glucuronidation capacity. Be­ Nielsen et al. (1992); however, it is contradictory to the cause the subjects were all Japanese, we could not tell fact that the quarternary ammonium-linked glucuron­ whether such an interethnic difference exists or not. ide of amitriptyline has been found in urine from pa­ However, interethnic difference in glucuronidation ca­ tients receiving therapeutic doses of amitriptyline, pacity has been reported using other drugs than TCAs which suggests direct conjugation of the parent com­ (Yue et al. 1989). pound, amitriptyline (Dahl-Puustinen et al. 1987, 1989; Lehman et al. 1983). Nielsen et al. (1992) reported that incubation of sam­ Impact of Administration of Age, Nutritional ples from Caucasian subjects who were treated with State, Benzodiazepine Derivatives, Smoking, clomipramine hydrochloride with B-glucuronidase/ or Gender on Desmethylation, Hydroxylation arylsulfatase increased the plasma concentrations of HC of C, and Glucuronidation of Hydroxylated and HDC by average factors of about 7 and 3, respec­ Metabolites of C tively, while Suftin et al. (1988) reported that apprecia­ ble glucuronide conjugate accumulation was observed No significant relationships were shown between age in patients treated with 200 mg/day of imipramine hy­ and individual capacities of desmethylation, hydroxy­ drochloride with average serum concentration ratios of lation, or glucuronidation of clomipramine. An increase 8.1 and 6.2 for 2-hydroxyimipramine glucuronide/2- in age has been reported to be associated with increased hydroxyimipramine and 2-hydroxydesipramine gluc­ plasma levels of TCAs and their metabolites (Bertils­ uronide/2-hydroxydesipramine, respectively. Bock et son et al. 1979; Kitanaka et al. 1982; Nies et al. 1977). al. (1982) reported an average serum concentration ra­ However, positive association between age and plasma tio ( = 2-hydroxydesipramine glucuronide/2-hydroxy­ levels of TCAs and its metabolites is generally attrib­ desipramine) of approximately 12 in plasma from sub­ uted to decreased renal clearance (Bertilsson et al. 1979; jects treated with 2.5 mg/kg/day of desipramine, while Cutler et al. 1981; Kitanaka et al. 1982; Nies et al. 1977), Bock et al. (1983) and Lieberman et al. (1985) reported but not to decreased hepatic enzyme activity. average ratio [ = (unconjugated 10-hydroxyamitript­ No significant relationships were shown between yline + unconjugated 10-hydroxynortriptyline)/(10- individual capacities of desmethylation, hydroxylation hydroxyamitriptyline glucuronide + 10-hydroxynor­ of clomipramine, or glucuronidation of hydroxylated triptyline glucuronide) of about 4 and 1, respectively. metabolites of clomipramine and nutritional state in the Deconjugation of our samples increased the plasma present study, but nutritional state affects the glucuroni­ concentrations of HC and HDC by average factors of dation of certain kinds of drugs [e.g., fenofibrate (Liu 3.36 and 3.61, respectively (i.e., HCG/HC = 2.36 and et al. 1991)]. NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4 Clomipramine Metabolism in an Asian Population 331

No significant effect of administration of benzodi­ ipramine (Brqisen et al. 1986a) and 20-fold interin­ azepine derivatives on desmethylation and hydroxy­ dividual variations for amitriptyline (Bock et al. 1982), lation of clomipramine was detected in the present respectively, while metabolic ratios for hydroxylation study, which seems to be consistent with previous of desipramine had approximately 100-, 102-, and 150- reports (Bertilsson et al. 1988; Otani et al. 1987; Trask­ fold interindividual variations (Bock et al. 1983; Potter man et al. 1979). However, a more controlled study is et al. 1982; Spina et al. 1984). Also Balant-Georgia et necessary to clarify the matter because a few of the many al. (1986) analyzed the data on steady-state concentra­ kinds ofbenzodiazepines used in this study should not tions of clomipramine and its hydroxylated and des­ have been administered as they can be suspected to methylated metabolites and reported approximately 80- have interaction with metabolism of tricyclic antidepres­ fold interindividual variations of hydroxylation of sants. clomipramine. A significant effect of benzodiazepine administra­ Reference to previous reports using such probe tion on glucuronidation was observed, but no previ­ drugs as sparteine, debrisoquine, or mephenytoin in ous studies that include analyses of the effect of ben­ a Japanese population might be useful in interpreting zodiazepine administration on glucuronidation of TCAs our data on desmethylation and hydroxylation. Al­ have been reported so far. though Nakamura et al. (1985) reported no poor Although a significant effect of smoking on des­ metabolizers of debrisoquine in 100 Japanese subjects, methylation and hydroxylation of clomipramine was Ishizaki et al. (1987) reported an incidence of poor shown in this study, this observation does not con­ metabolizers of sparteine as 2% in 84 Japanese subjects. tradict the fact that smoking induces some cytochrome Nakamura et al. (1985) also reported that the incidence enzymes (Jusko 1979). of poor metabolizers of mephenytoin was 18%, while Significant gender effects on desmethylation and Jurima et al. (1985) found that 7 of 31 Canadians of Jap­ hydroxylation were observed, with female subjects hav­ anese descent (incidence of 23%) were poor mepheny­ ing a lower glucuronidation capacity of hydroxylated toin metabolizers. From the results of probit analyses metabolites of clomipramine than did male subjects in of MRD and MRH, no apparent poor desmethylators the present study (GR: male vs. female 0.37 vs. 0.64). or poor hydroxylators were observed in our 108 sub­ Although Someya et al. (1992) reported the extent of jects. Our data obtained on hydroxylation of C appeared glucuronidation of haloperidol and reduced haloper­ to be consistent with the result of Ishizaki et al. (1987), idol, further analysis of their data revealed that but our data on desmethylation contradict those of both glucuronidation capacity of female subjects was approx­ Jurima et al. (1985) and Nakamura et al. (1985). imately 60% of those of male subjects, which is com­ Four possible explanations can be made for these patible with the result of the present study. discrepancies. First, such probe drug studies were per­ formed in the single-dose study design, while our data was obtained during chronic treatment of C. Second, Interindividual Variations of Desmethylation chronic administration of clomipramine may induce the and Hydroxylation of Clomipramine enzymes responsible for desmethylation and hydroxy­ Reference to several reports on the metabolism of TCAs lation for C and phenotypes of poor hydroxylators or in vivo provided controversial findings. Although meta­ desmethylators of clomipramine are masked. Third, a bolic ratios for desmethylation corresponding to MRD period of 2 weeks is too short to obtain the steady state, in this study for various TCAs showed 9- to 15-fold vari­ particularly in poor metabolizers. Fourth, the enzymes ations for imipramine (Bnt,sen et al. 1986a, 1986b), responsible for desmethylation of TCAs may not be 5-fold variations for clomipramine (Traskman et al. identical to mephenytoin oxygenase (Gram et al. 1990; 1979), and 12- to 14-fold interindividual variations for Skjelbo et al. 1991). amitriptyline (Bock et al. 1982; Breyer-Pfaff et al. 1982), respectively, Gram et al. (1977) and Kuss and Jungkunz Interindividual Variation of Glucuronidation of (1986) reported approximately 80-fold interindividu­ Hydroxylated Metabolites of C al variations of desmethylation of imipramine and clomipramine, respectively. Approximately 36-fold in­ Approximately 28-fold interindividual variation of GR, terindividual variations in MRD were observed in our ranging from 0.13 to 3.69 were observed in 108 subjects study. investigated in this study. Reference to the previous report on the metabolism Few studies on interindividual variation of gluc­ of TCAs also suggested controversial findings for meta­ uronidation of TCAs has been reported so far. How­ bolic ratios for hydroxylation: metabolic ratios for ever, Bock et al. reported an about fourfold interin­ hydroxylation calculated from steady-state concentra­ dividual variation of GR in 27 patients receiving tions exhibited 24-fold interindividual variations for im- amitriptyline (Bock et al. 1982), and Suftin et al. (1988) 332 K. Shimada et al. NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4

reported an approximately 13-fold interindividual vari­ Breyer-Pfaff U, Gaertner HJ, Kreuter F, Scharek G, Brink­ ation in GR in 14 subjects receiving 200 mg of imipra­ schulte M, Wiatr R (1982): Antidepressant effect and phar­ mine for 7 days. Interindividual variation in GR ob­ macokinetics of amitriptyline with consideration of un­ bound drug and 10-hydroxynortryptyline plasma levels. served in this study seems to be larger than those of Psychopharmacol 76:240-244 previous reports (Bock et al. 1982; Suftin et al. 1988), Bnt,sen K, Gram LF, Klysner R, Bech P (1986a): Steady state and this might be due to the difference in sample size levels of imipramine and its metabolites: Significance of between previous reports and ours. dose-dependent kinetics. Eur J Clin Pharmacol 30:43-49 Bnt,sen K, Klysner R, Gram LF, Otton SV, Bech P, Bertilsson L (1986b): Steady-state concentrations of imipramine and ACKNOWLEDGMENTS its metabolites in relation to sparteine/debrisoquine poly­ morphism. Eur J Clin Pharmacol 30:679-684 The authors express their gratitude to Drs. J. Szeszak and K. Cutler NR, Zavadil AP, Eisdorfer C, Ross RJ, Potter WZ (1981): Scheibli, Ciba-Geigy Limited, Basel, Switzerland, for provid­ Concentrations of desipramine in elderly women are not ing N-desmethylclomipramine hydrochloride, 8-hydroxy­ elevated. Am J Psychiatry 138:1235-1237 clomipramine base, 8-hydroxy-N-desmethylclomipramine hy­ Dahl-Puustinen ML, Bertilsson L (1987): Formation of a quar­ drochloride; to Ciba-Geigy Japan Limited for providing ternary N-glucuronide of amitriptyline in human liver clomipramine hydrochloride; and to Dr. Koichi Otani (Depart­ microsomes. Pharmacol Toxicol 61:342-346 ment of Psychiatry, Hirosaki University, Hirosaki, Japan) for collecting samples and for his invaluable advice and sugges­ Dahl-Puustinen ML, Aberg-Wistedt A, Bertilsson L (1989): tions in preparing the manuscript. This research was partly Glucuronidation of amitriptyline in man in vivo. Phar­ supported by a grant-in-aid from the Japanese Ministry of macol Toxicol 65:37-39 Health and Welfare, the Japanese Ministry of Education, and Daniel W, Melzacka M (1986): The effect of neuroleptics on a grant from the Pharmacopsychiatry Research Foundation. imipramine desmethylation in rat liver microsomes and imipramine and desipramine level in the rat brain. Bio­ chem Pharmacol 35:3249-3253 Della Corte L, Broadhurst AD, Sgaragli GD, Fillippini S, Hee­ REFERENCES ley AF, James HD, Favarelli C, Pazzagli A (1979): Clini­ cal response to tricyclic plasma levels during treatment Alexanderson B, Price Evans DA, Sjoqvist F (1969): Steady­ with clomipramine. Br J Psychiatry 134:390-400 state plasma levels of nortriptyline in twins: Influence Edelbroek PM, Zitman FG, Knappert-van der Klein EAM, of genetic factors and drug therapy. Brit Med J 4:764-768 van Putten PM, de Wolff FA (1987): Therapeutic drug American Psychiatry Association (1987): Diagnostic and monitoring of amitriptyline: Impact of age smoking and statistical manual of mental disorders, ed 3, rev., contraceptives on drug and metabolite levels in bulimic Washington, DC, American Psychiatry Association woman. Clin Chim Acta 165:177-187 Balant-Gorgia AE, Schulz P, Dayer P, Balant L, Kubli A, Fekete J, Del Castilho P, KraakJC (1981): Reversed-phase liq­ Gertsch C, Garrone G (1982): Role of oxidation polymor­ uid chromatography for the separation of chlorproma­ phism on blood and urine concentrations of amitripty­ zine, imipramine and some of their metabolites. J Chro­ line and its metabolites in man. Arch Psychiatr Nervenkr matogr 204:319-327 232:215-222 Gram LF, S(/Jndergaard I, Christiansen J, Petersen GO, Bech Balant-Gorgia AE, Balant LP,Genet C, Dayer P, Aeschlimann JP, Reisby N, Ibsen L, Ortmann J, Nagy A, Dencker SJ, JM, Garrone G (1986): Importance of oxidative polymor­ Jacobsen 0, Krautwald O (1977): Steady-state kinetics phism and levomepromazine treatment on the steady­ of imipramine in patient. Psychopharmacology (Berlin) state blood concentrations of clomipramine and its ma­ 54:255-261 jor metabolites. Eur J Clin Pharmacol 31:449-455 Gram LF, Br(/Jsen K, Skjelbo E, Sidrup S (1990): Significance Bertilsson L, Aberg-Wistedt A (1983): The debrisoquine of genetic polymorphism and selective P450 interactions hydroxylation test predicts steady state plasma levels of in psychopharmacology. Clinical Neuropharmacology desipramine. Br J Clin Pharmacol 15:388-390 13(suppl 2):559-560 Bertilsson L, Mellstrom B, Sjoqvist F (1979): Pronounced in­ Hammer W, Sjoqvist F (1967) Plasma levels of monometh­ hibition of noradrenaline uptake by 10-hydroxy­ ylated tricyclic antidepressants during treatment with metabolites of nortriptyline. Life Sci 25:1285-1292 imipramine-like compounds. Life Sci 6:1895-1903 Bertilsson L, Aberg-Wistedt A, Liden A, Otani K, Spina E Ishizaki T, Eichelbaum M, Horai Y, Hashimoto K, Chiba K, (1988): Alprazolam does not inhibit the metabolism of Dengler HJ (1987): Evidence for polymorphic oxidation nortriptyline in depressed patients or inhibit the metab­ of sparteine in Japanese subjects. Br J Clin Pharmacol olism of desipramine in human liver microsomes. Ther 23:482-485 Drug Monit 10:231-233 Jusko WJ (1979): Influence of cigarette smoking on drug me­ Bock JL, Giller E, Gray S, Jatlow P (1982): Steady-state plasma tabolism in man. Drug Metab Rev 9:221-228 concentrations of cis- and trans-10-OH amitriptyline me­ Jurima M, Inaba T, Kadar D, Kalow W (1985): Genetic poly­ tabolites. Clin Pharmacol Ther 31:609-616 morphism of mephenytoin p(4')-hydroxylation: Differ­ ence between Orientals and Caucasians. Br J Clin Phar­ Bock JL, Nelson JC, Gray S, Jatlow PI (1983): Desipramine macol 19:483-487 hydroxylation: Variability and effect of antipsychotic drugs. Clin Pharmacol Ther 33:322-328 Jacqz E, Hall SD, Branch RA (1986): Genetically determined NEUROPSYCHOPHARMACOLOGY 1995-VOL. 12, NO. 4 Clomipramine Metabolism in an Asian Population 333

polymorphism in drug oxidation. Hepatology 6:1020- Nordin C, Swiers B, Benitez J, Bertilsson L (1985): Plasma con­ 1032 centrations of nortriptyline and its 10-hydroxy metabo­ Kitanaka I, Ross RJ, Cutler NR, Zavadil AP, Potter WZ (1982): lite in depressed patients-relationship to the debriso­ ratio. Br Clin Pharmacol Altered hydroxydesipramine concentrations in elderly quine hydroxylation metabolic J depressed patients. Clin Pharmacol Ther 31:51-55 19:832-835 L (1987): No interaction of di­ Kruger R, H6zl G, Kuss HJ, Schefold L (1986): Influence of Otani K, Nordin C, Bertilsson neuroleptics on the metabolism of tricyclic antidepres­ azepam on amitriptyline disposition in depressed pa­ sants-In vitro experiments with rat liver microsomes. tients. Ther Drug Monit 9:120-122 Biol Psychiatry 21:1247-1257 Potter WZ, Calil HM, Suftin TA, Zavadil AP, Jusko WJ, (1982): Active metabolites of Kuss HJ, Jungkunz G (1986): Nonliner pharmcokinetics of Rapoport J, Goodwin FK chlorimipramine after infusion and oral administration imipramine and desipramine in man. Clin Pharmacol in patients. Prog Neuro-Psychopharmacol Biol Psychiatry Tuer 31:393-401 10:739-748 Siest G, Antonio B, Fouurnel S, Magdalow I, Thomassin I (1987): The glucuronyltransferases: What progress can Lehman JP,_ Fen~elau C, Depaulo JR (1983): Quarternary pharmacologists expect from molecular biology and cel­ a°:1momum-linked glucuronides of amitriptyline, imipra­ 36:983-989 mme, and chlorpromazine. Drug Metab Dispos 11: lular enzymology? Biochem Pharmacol 221-225 Skjelbo E, Br6sen K, Hallas J, Gram LF (1991): The mepheny­ is partially responsible for Lieberman JA, Cooper TB, Suckow RF, Steinberg H, Boren­ toin oxidation polymorphism the N-demethylation of Imipramine. Oin Pharmacol Tuer stein M, Brenner R, Kane JM (1985): Tricyclic antidepres­ sant and metabolite levels in chronic renal failure. Clin 49:18-23 Pharmacol Ther 37:301-307 Someya T, Shibasaki M, Noguchi T, Takahashi S, Inaba T in psychiatric patients: Liu HF, Vinvent-ViryM, GalteauMM, GueguenR, Magdalou (1992): Haloperidol metabolism of glucuronidation and carbonyl reduction. J, Nicolas A, Leroy A, Siest G (1991): Urinary glucuronide Importance excretion of fenof:tbric and clof:tbric acid glucuronide in J Clin Psychopharmacol 12:169-174 man. ls it polymorphic? Eur J Clin Pharmacol 41: 153-159 Spina E, Birgersson C, von Bahr C, Ericsson 6, Mellstr6m (1984): Pheontypic consistency Mellstr6m B, Bertilsson L, Sa.we J, Sch4lz HU, Sjoqvist F B, Steiner E, Sj6qvist F (1981): E- and Z-10-hydroxylation of nortriptyline: Rela­ in hydroxylation of desmethylimipramine and debriso­ tionship to polymorphic debrisoquine hydroxylation. quine in healthy subjects and in human liver microsomes. Clin Pharmacol Ther 30:189-193 Clin Pharmacol Ther 36:677-682. Spina E, Steiner E, Ericsson Sj6qvist F (1987): Hydroxyla­ Mellstro~ ~' B~rtilsson L, Lou Y-C, Sa.we J, Sjoqvist F (1983): 6, Dependence on the Am1~npty~me metabolis~: Relationship to polymorphic tion of desmethylimipramine: phenotype. Clin Pharmacol debnsoqume hydroxylation. Clin Pharmacol Ther 34: debrisoquine hydroxylation 516-520 Ther 41:314-319 Mellstrom B, Sa.we J, Bertilsson L, Sj6qvist F (1986): Amitrip­ Suftin TA, Perini GI, Molnar G, Jusko WJ (1988): Multiple­ tyline metabolism: Association with debrisoquin hy­ dose pharmacokinetics of imipramine and its major ac­ droxylation in non-smoker. Clin Pharmacol Ther 39: tive and conjugated metabolites in depressed patients. 369-371 J Clin Psychopharmacol 8:48-53 Nakamura K, Goto F, Ray WA, McAllister CB, Jacqz E, Wilkin­ Traskman L, Bertilsson L, Asberg M, Cronholm B, Mellstrom son GR, Branch RA (1985): Interethnic differences in B, Neckers LM, Sjoqvist FE, Thoren P, Tybring G (1979): genetic polymorphism of debrisoquin and mephenytoin Plasma levels of chlorimipramine and its demethyl me­ hydroxylation between Japanese and Caucasian popu­ tabolites during treatment of depression. Oin Pharmacol lations. Clin Pharmacol Ther 38:402-408 Ther 26:600-610 Nielsen KK, Brq,sen K, Gram LF, The Danish University An­ tidepressant Group (1992): Steady-state plasma levels of Woolhouse NM, Adjepon-Yamoah KK, Mellstrom B, Hed­ clomipramine and its metabolites: Impact of the spar­ man A, Bertilsson L, Sjoqvist F (1984): Nortriptyline and teine/debrisoquine oxidation polymorphism. Eur J Clin debrisoquine hydroxylations in Ghananian and Swed­ Pharmacol 43:405-411 ish subjects. Clin Pharmacol Ther 36:374-378 Nies A, Robinson DS, Friedman MJ, Green R, Cooper TB, Yue QY, Svensson JO, Alm C, Sjoqvist F, Sa.we J (1989): In­ Ravari_s CL, _Ives JO (1977): Relationship between age and terindividual and interethnic differences in the demethyl­ tncychc antidepressants plasma levels. Am J Psychiatry ation and glucuronidation of codeine. Br J Oin Pharmacol 134:790-793 28:629-637